Prediction of Deceleration Amount of Vehicle Speed in Snowy Urban Roads Using Weather Information and Traffic Data

In snowy countries, heavy snow has a large influence on traffic flows. Snow on urban roads disturbs traveling of vehicles as a huge amount of snow is piled up on roadsides, which often obstructs smooth driving. In this paper, we propose a novel method to predict the speed of vehicles on each road segment in snowy cities. This estimation is helpful for urban traffic planning of local government or trip planning of residents. We collect weather information such as the highest temperature, daylight hours, snow depth and snowfall of the previous day and current new snowfall and vehicular traffic data of each road segment obtained from floating car data such as the average speed in summer and the speed of the previous day. We have built a speed model of vehicles for each road segment in snowy conditions as a linear combination of those factors. Then we enumerate multiple factors which might have some influence on vehicle speed in snowy conditions and have derived their weights by using multiple regression analysis. We have applied the proposed method to major road segments in Sapporo City, Japan and derived multiple regression functions using real weather information and vehicular traffic data. We have shown that our proposed model can predict the speed deceleration for seven days with small errors.